Research On Short-term Power Consumption Prediction Based On Deep Neural Network

Electric energy plays a pivotal role in the production and life of the people.Compared with other energy sources,electric energy has the characteristics of balancing power transmission,distribution,and difficulty in large-scale storage.In order to achieve instant balance of power demand and effectively avoid waste of electric energy,Need to make short-term forecasts of power consumption.The short-term forecast of the power consumption of the power system can not only meet the power demand of the power users at the load side,but also provide accurate power demand for the power system marketers participating in the power system marketing,thereby ensuring that the power marketers target the power users.Demand for electricity sets competitive electricity prices.Therefore,it is of great practical significance to conduct research on short-term power consumption forecasting of power systems.In this paper,a deep neural network is used to carry out short-term power consumption forecasting research in the power system.Based on the construction of a deep neural network model,experimental simulation verification of related algorithms is performed based on actual power consumption data in a certain area.First,based on the actual power consumption data of a certain area,carry out the analysis of the relationship between the daily cycle characteristics of the actual power consumption in the area and the meteorological factors and the user's power consumption characteristics under different periods;second,the actual power consumption Based on the analysis of electricity data,based on the neighborhood rough set theory,carry out feature extraction research of influencing factors of electricity consumption,and make predictions based on the extracted features;Then,aiming at the shortcomings of traditional BP neural network,such as slow convergence and easy to fall into local optimum,this paper proposes a BP neural network(BABP)optimized using a bat algorithm and a prediction algorithm based on a deep confidence network(DBN).The best network structure for DBN.Final,based on BP neural network(BABP)and Deep Belief Network(DBN)two models for short term power consumption prediction research,the analysis of experimental results shows that both models are suitable for training a large number of samples,and the prediction accuracy of the DBN model is relatively High and best generalization performance.Then,the phase error index is analyzed with a specific example,and the DM test method is used to verify that the short-term power consumption prediction model based on the deep confidence network has better prediction ability.